This research seeks to test a causal model of turnover. The type of turnover to be studied is voluntary separations. A review of the literature and a pilot study has produced a list of fourteen causally-significant variables to be tested. Most of the data for this research have already been collected by means of a grant from the American Nurses' Foundation. A total of 1077 nonsupervisory, registered nurses completed a mail questionnaire in August 1976 for a response rate of seventy-seven percent. These nurses are employed by seven, voluntary, short-term hospitals located in eastern Iowa and western Illinois. Data about the actual turnover of the nurses will be collected from the hospitals in December 1977. Funds are primarily requested to analyze the data and write up the results. The data will be analyzed by means of multiple regression analysis. Funds are also requested to communicate the results of the research to interested parties in the health-delivery system. Two years are needed for the analysis, write up, and communication of the results.